On August 5, chief AI scientist Jeff Dean announced he was leaving Google after 27 years. Google proceeded to lose ~$180 billion in market cap, making this the largest impact I've ever seen by one employee departing.
He's co-founding Discovery Loop, a public benefit corporation aimed at using AI to automate the scientific research process itself. He's bringing Sanjay Ghemawat, Oriol Vinyals, and Quoc Le with him. Alphabet is a founding investor and cloud partner.
That is the most decorated founding team in the history of applied machine learning. Every venture firm on Sand Hill Road would take that meeting and commit on 30 minutes’ notice at likely any valuation.
Here's the thing. If you run a $100 million venture fund or less, you are not in that round. You are not even in the building. There is also no chance of you as a retail investor, investing in Jeff's Discovery Loop either, unless you commit capital to a venture mega-fund ($1+ billion) as a limited partner.
I want to explain why, because it changes how you should think about investing in future venture capital funds for private company exposure.
The Question Nobody Asks
I wrote about the SpaceX lockup expiration and my modest little distribution of 92 shares, 20% of my holdings.
The fund bought SpaceX at $16.20 a share. Shares were distributed against a reference price of $133.11. That's an 8.2X multiple on invested capital in two years and nine months.
Nice return. But the more interesting question isn't what the fund made. It's how the fund got in at all.
In late 2023, SpaceX was valued in the ballpark of $150 billion. At that scale, a company isn't taking a $1 million check. It probably isn't taking a $5 million check either. Practically speaking, you need to write $10 million or more to be worth the paperwork and the line on the cap table.
So run the math backwards.
If you want a $10 million position to be a meaningful 5% of your fund, you need at least a $200 million fund. But most venture funds hold 30 to 50 positions, not 20. Therefore, to build such a portfolio and still write $10 million into a late-stage winner, you’re looking at $335 million in assets minimum.
My 8.2 MOIC exists because somebody raised a fund large enough to be allowed into the room.
Large Venture Funds It Isn't Only About Fees
The cynical, but still truthful take is that managers of mega-funds want bigger fees. On 2% to 2.5% management and 20% to 35% carry, a bigger fund does mean a bigger paycheck regardless of performance. Being a venture capitalist at a large fund is one of the most lucrative occupations in the world.
But the optimist believes the desire for more money isn't the main driver.
Funds got bigger because the companies got bigger and stayed private longer. SpaceX waited 20 years to go public. Stripe, Databricks, OpenAI, and Anthropic have all raised at valuations that would have been considered late-stage public market territory a generation ago.
If the best companies aren't going public and the private rounds are enormous, a small fund doesn't get to participate in the upside at all. Sequoia, Thrive, ICONIQ, and a16z raise billions because the alternative is watching the best deals happen without them.
Bigger funds aren't a symptom of greed. They're a symptom of companies staying private longer and growing much larger.
Some Seed Rounds Are Now Massive
Here's how far this has traveled down the stack.
Go back to Jeff Dean. The Discovery Loop valuation hasn’t been publicly disclosed. But when a team like that raises a first round, “seed” is a formality.
I was guessing Discovery Loop might raise at a $2 billion valuation compared to YC companies raising at a $30-$40 million valuations. I was wrong by 5X. Business Insider is reporting that Jeff Dean's Discovery Loop is looking to raise $1 billion at a $10 billion valuation! This one of the largest, if not the largest seed valuations ever.
If you want to co-lead a $1 billion funding round, then your fund likely has to invest around $100 million. To participate, I'm not sure Jeff and company would bother accepting any check under $10 million, unless it was from friends and family.
If your entire fund is $100 million or even a highly respectable $350 million, you're not leading that. I'm not sure your fund is even getting on the cap table. Companies want a tight cap table of investors who can support them through multiple additional rounds. They can tell in one meeting whether you have the balance sheet to follow on, and they'd rather have five investors who can support their position than twenty who can't.
The word “seed” now describes a stage, not a size.
The Two-Fund Venture System
Which is why the top firms almost always raise an early-stage fund and a growth fund in the same vintage year.
The early fund finds the diamond. The growth fund defends the ownership percentage when the diamond gets expensive.
Without the second fund, you get diluted out of your best position precisely when it starts working. You found the company at $50 million and watched your 8% shrink to 3% by the time it's worth $10 billion. All the conviction, a fraction of the return.
Two funds, one machine. If you're evaluating a manager and they only have one of the two, ask what happens to their winners.
The Venture Stage I Mostly Prefer To Invest
Given a FIRE investor can't afford to be too wrong, I prefer later-stage growth funds, meaning Series C and beyond with most of my venture money.
The return multiples are lower because the company already has product-market fit and meaningful revenue. But the probability of getting my money back is meaningfully higher.
At 49, with no paycheck and two kids under 10, I'll take lower returns with greater certainty. Compounding a 3X on capital I'm confident I'll see again is more attractive than a lottery ticket on a 20X I might not.
Take Anthropic. I would have happily invested in May 2026 at a $965 billion post-money valuation. Annualized revenue had crossed $47 billion, and backers now expect $100 billion to $120 billion by December 2026. This week the Financial Times reported investors expect an October IPO at $2 trillion or more, which would be the largest listing in history. Even bigger than SpaceX.
Worth noting that the $2 trillion figure comes from investors building their own models, not from Anthropic. Executives reportedly haven't set a target even in private. But if it lands anywhere close, that's a double in five months on a company already valued near a trillion dollars.
So could I have invested in May? Nope. The firms in that round were Altimeter, Dragoneer, Greenoaks, Sequoia, Capital Group, Coatue, D1, GIC, ICONIQ, and XN. Every one runs multi-billion dollar funds. If you don't have a fund that size, or you're not an LP in one of them, you're out of luck.
Which is why I'm thankful more than 20% of VCX sits in Anthropic. If the IPO prices anywhere near $2 trillion, my VCX NAV estimate has meaningful upside. Let's see if the market finally starts paying attention to my analysis.
That said, early-stage is where the explosive upside lives. Seed, Series A, Series B. That's where a 20X or 100X actually happens. If you're 30 years old with 30 years of earning power ahead of you, weight yourself differently than I do. Your ability to earn your way out of a mistake is your highest-yielding asset.
I wrote about that tradeoff in The Problem With Investing In Venture Capital When Older.
Fund Size Isn't The Only Moat
Luckily, if you're interested in investing in smaller companies, money is no longer differentiated. Every large fund has money by definition. If a founder is choosing between two billion-dollar funds writing identical checks at identical valuations, the money is not the variable.
What's scarce now is the ability to put a company in front of people who will actually use it. I made this argument in Distribution May Be The Last Moat. AI has made building cheap and marketing noisy. The bottleneck moved to attention.
You can see firms responding in real time. Lightspeed hired Claire Zau, a creator with over 250,000 Instagram followers and 100,000 on TikTok, as Partner of New Media. She sources deals and co-hosts the firm's new show, Lightwork. a16z bought Erik Torenberg's Turpentine podcast and made him a general partner. OpenAI acquired TBPN. Half of Sand Hill Road now runs a newsletter or a podcast.
Firms are recognizing capital alone is a commodity. To win the best deals you need a personality and real relationships, not just a wire transfer.
Maybe I should get into this venture capital business as well.
A Case Study In Capital Plus Distribution
Fundrise's Innovation Fund (NYSE: VCX) is an interesting example of both variables at once.
The fund reported net assets of roughly $679 million as of March 31, 2026, with top holdings including Anthropic, Databricks, and OpenAI. Its top ten positions made up about 67% of net assets. Actual AUM today is likely well above that, since my NAV estimate is at least 60% higher than the March mark. So it's a real fund by size, comfortably able to write institutional checks.
But the more interesting asset isn't the balance sheet. Fundrise has hundreds of thousands of investors on its platform and an email list several times larger.
Think about what that's worth to a portfolio company. A consumer fintech that lands Fundrise on its cap table gets a shareholder who can introduce it to hundreds of thousands of financially engaged people in a single send. Not a logo on a website. Actual users.
A Sand Hill Road firm can make warm intros to enterprise buyers and recruit you a VP of Engineering. Clearly valuable. But it cannot hand you hundreds of thousands of retail consumers who already trust the brand.
Full disclosure: I've been a Fundrise affiliate partner since 2016 and I hold VCX shares I bought before it listed.
What This Means If You're Not Writing $10 Million Checks
Almost nobody reading this is getting allocated into a Discovery Loop's ~$1 billion seed round. Radical Ventures and Khosla Ventures are co-leading, with Lightspeed, Kleiner Perkins, Doerr Capital, and Alphabet participating. Google is both a founding investor and the cloud partner, supplying compute for at least the first year.
The round was still open at announcement and the valuation remains undisclosed, which tells you it isn't being marketed to anyone who has to ask.
That table was set before most people had read the news.
But the takeaway isn't defeatist. Access to private companies now runs through vehicles rather than relationships. Either you commit to a fund large enough to be in the room, or you buy exposure through something publicly traded that owns the assets.
Ten years ago the second option barely existed. That's the actual democratization story, and it gets buried under complaints about fund sizes.
The funds that will win the next decade are the ones with capital and an audience.
Reader Questions
If you're an LP in venture funds, has fund size changed your allocation decisions? Are you going bigger with fewer managers, or the opposite?
Do you think distribution has genuinely become the scarce asset in venture, or is that just what people with audiences want to be true?
Understanding how capital compounds at each stage of wealth is the difference between a nice line item and a new chapter. I wrote Millionaire Milestones: Simple Steps To Seven Figures, a USA Today bestseller published by Portfolio Penguin Random House, to lay out that path step by step.
Subscribe To Financial Samurai
To get my posts in your inbox, join 60,000+ readers and sign up for the free Financial Samurai newsletter here. For new post notifications, ad-free for the first few hours, sign up here.
Money is too important to be left up to pontification, which is why I've been writing from firsthand experience since 2009.
